3 ASSESSMENT OF MICROALGAE BASED PRODUCTION PROCESSES
Snook (1983). The β-carotene content of the cells was expected to be 5 wt% of the cell dry weight (dw), similar to the extraction results of Dunaliella biomass from the open pond plant in Western Australia and India (Sources of biomass: Nutra-Kol and Denk Ingredients GmbH). To achieve optimal growth conditions, the raceway pond is actively fed with CO2 and nutrients, whereas the CO2 and nutrient consumption were calculated based on the stoichiometric equation of algae biomass according to Bailey Green et al. (1996): C106H181O45N16P. The demand of the nutrients (NH4)3PO4
and NH3 was calculated considering an excess of 10% for the blow-down and loss of nutrients further downstream. Under the assumption of a complete medium recycling after dewatering, a reusability of 5% nutrients was expected. The CO2 uptake in the algae was expected to be 85% eective. CO2 delivery occurs with a blower consuming 0.025 kWh m−3 (Verrecht et al., 2010). The daily energy consumption of the blower was estimated considering a CO2 density of 2.23 kg m−3 at an average process temperature of 19.8◦C and a delivery pressure of 1.22 bar, as proposed by Lundquist et al. (2010).
3.2.2 Water supply and pumping
Recycled water from the harvesting step was supplied for medium preparation consid-ering a blow-down of 5%. For the compensation of blow-down and evaporation losses, make-up water was assumed to be taken from the nearby sea. Furthermore, natural evaporation and rainfall were taken into account using weather data from the existing D. salina production site in Hutt Lagoon (Australia) to simulate realistic conditions.
For this purpose, evaporation and rainfall values of 2445 mm a−1 and 449.7 mm a−1 (Nowicki et al., 2009) were applied to estimate the average net evaporation rate of 5.5 mm d−1. The cost of sea water cleaning by ltration and chlorination was set to 0.54 USD m−2according to data from the NBT Ltd. plant in Israel (Ben-Amotz, 2008).
Pumping work of recycled and make-up water as well as the transport of microalgae suspension through the process units was supposed to be done by conventional waste water pumps. Therefore, an energy demand of 0.016 kWh m−3 was used, adopted from the work of Verrecht et al. (2010).
3.2.3 Dewatering procedure
The downstream route typically found in microalgae process analysis is composed of a multi-step harvesting procedure (Davis et al., 2011; Delrue et al., 2012; Sun et al., 2011;
Weschler et al., 2014). Most commonly, dewatering is initiated by a simple sedimenta-tion step without the addisedimenta-tion of occulants. As demonstrated in the preliminary results obtained in this thesis, D. salina has an insucient settling eciency of less than 40%
within one day (see Appendix Figure A.1). Thus, this subunit was not implemented and centrifugation was directly used for cell harvesting in the simulated process. Therefore, a daily harvesting volume of 667 m3 was supposed to be pumped into the centrifuge.
This volume corresponds to the daily biomass yield of 200 kg (see Section 3.2.1).
A harvesting eciency ηH of 95% was used for the centrifugation step, provided by Prof. Ami Amotz from NBT Ltd. Israel (personal communication with Prof.
Ben-3.2 Development of a process scheme for microalgal β-carotene production
Table 3.1: Parameters incorporated in the process model.
Parameter Value Unit
Location data
Man power cost 5,000 USD ha−1 a−1
Domestic tax 50,000 USD ha−1 a−1
Temperature 19.8 ◦C
Cultivation period 330 d a−1
Pond depth 0.2 m
Production area 10 ha
Evaporation 2445 mm a−1
Rainfall 449.7 mm a−1
Water treatment cost 0.54 USD m−2
Cultivation
Species CCAP 19/18
Growth rate 2 gdw m−2 d−1
β-carotene content 5 %
Cell concentration 0.3 gdw L−1
Nutrient blow down 10 %
Water blow down 5 %
Nutrient recycle 5 %
CO2uptake eciency 85 %
NaCl concentration 24 %
Harvesting volume 667 m3d−1
Downstream processing
Harvesting eciency 95 %
Post-harvest concentration 100 gdw L−1
Drying eciency 99 %
Post-drying concentration 900 gdw L−1
Extraction eciency 95 %
Loss of solvent 1 %
Equipment
Paddle wheel velocity 0.25 m s−1
Paddle wheel energy demand 574 kWh d−1
Blower energy consumption 0.025 kWh m−3
Pump energy consumption 0.016 kWh m−3
Centrifuge capacity 85 m3h−1
Centrifuge motor power 75 kW
Dryer energy consumption 1.31 kWh kg−1H2O
Extractor energy consumption 0.00035 kWh kg−1dw
Heater extraction energy consumption 1.5 kWh kg−1dw
3 ASSESSMENT OF MICROALGAE BASED PRODUCTION PROCESSES
Amotz, 2014). Consequently, 95% of the applied biomass obtained in the cultivation step can be recovered by centrifugation. The dewatering approach was supposed to reach a nal biomass concentration of 10 wt% which is a typical value in microalgae processes (Ben-Amotz, 2008; Molina Grima et al., 2003). The energy demand of the centrifuge was estimated from Equation 3.1 (adopted from Xu et al. (2011)) by using the technical data of the Westfalia separator SSE 400 with a capacity of 85 m3 h−1 and a motor power of 75 kW:
Ec= mdw·Pc
calgae·νc (3.1)
where Ec is the energy in kWh required by the centrifuge, mdw is the dry weight of D. salina needed to be processed in kg, calgae is the mass concentration of the algae slurry in kg m−3, νc is the capacity of the centrifuge in m3 h−1 and Pc is the motor power of the centrifuge in kW. As a last dewatering step, drying was expected to reach a nal biomass concentration of 90%. According to Ben-Amotz & Avron (1989), spray drying, freeze-drying and drum drying is appropriate for Dunaliella biomass, resulting in a comparable quality of the biomass powder and stability of the pigment. For this reason, a spray-drying unit similar to the plant in Israel (Ben-Amotz, 2008) was used for process simulation, consuming 1.31 kWh kg−1H2O (Baker & McKenzie, 2005). The loss of biomass during spray drying was assumed to be 1%.
3.2.4 Extraction
After the dewatering procedure, the main productβ-carotene can be extracted by dif-ferent ways (see Section 2.1.3). Conventional extraction with non-polar organic solvents such as n-hexane is feasible due to the high extraction eciencies of at least 95% (Ru-ane, 1974). Based on simulations of bio-oil extraction with n-hexane from microalgae biomass at 60◦C by Delrue et al. (2012), electric power requirements of 1.3 kWh kg−1dw were applied for the heater and 0.00035 kWh kg−1dw for the extractor, respectively. This also considers the energy demand of the heater needed for the recovery of the solvent.
Solvent recovery was assumed to be 99% eective. Adopted from the patent of Ruane (1974), a relation of 10:1 hexane:algae slurry was assumed for the solvent to biomass ratio.
3.2.5 Implementation and statistics
To assess the economics of the production route in the underlying study, the mass and energy ows were calculated using the mentioned parameters. Mass ows refer to the in and out ows of the process as schematically illustrated in Figure 3.2. Energy ows account for the direct energy consumption within the individual process units. All data above are derived from published works, own experimental results or personal commu-nications with experts in the eld. The developed process model was implemented in Matlab (MathWorks) and can be generally expressed by the following equations:
OC =X
i,k
P1,i(Fi,kin−X
l
Ri,k+l) +X
j,k
P2Ej,k+X
i,k
P3,iCi,k+X
i,k
P4,i(Wi,k−Ri,k) (3.2)
3.2 Development of a process scheme for microalgal β-carotene production
GR=X
i,k
P5,iFi,kout (3.3)
with OC as net operating costs. P1,i, P2, P3,i and P4,i are the costs of the raw materialsFi,kin, the consumed energyEj,k of utility j(e.g. pump, centrifuge, heater) the needed chemicals Ci,k as well as the waste and recycling streams Wi,k and Ri,k. The gross revenue GR considers the sales price P5,i of component i in the product stream Fi,kout. Index i refers to mass components in the ow (e.g. water, biomass, pigment), whereask accounts for the process steps.
𝑬𝒋,𝒌 𝑹𝒊,𝒌+𝒍
𝑭𝒊,𝒌𝒊𝒏
𝑾𝒊,𝒌
𝑭𝒊,𝒌𝒐𝒖𝒕 𝑪𝒊,𝒌
𝑹𝒊,𝒌
Figure 3.2: Schematic illustration of the mass and energy ows in the used process model.
For the consideration of uncertainties of all parameter values used in the energy and cost analysis, Monte Carlo simulations were applied using 5 x 105 independent normally distributed samples. The variances were dened in dependence of the individual pa-rameter sources. In particular, a variance of σparameter2 = (0.253 µparameter)2 was applied for parameters derived from literature which are three standard deviations correspond-ing to 25% of the nominal parameter. A variance of σparameter2 = (0.13 µparameter)2 was used for parameters which are specic for D. salina. For the determined experimental parameters, the actual observed experimental variances were used. Thus, the sampling intervals with their specic upper and lower limits of each parameter were dened.
3.2.6 Reliability of the process model
The process model depicted in Figure 2.1 was used as reference case of subsequent investigations. The results of the process analysis clearly demonstrate the reliabil-ity of the used assumptions in the developed model. To be more precise, the esti-mated total annual process costs of 181,603 ± 19,330 USD a−1 are nearly identical to the 180,000 USD a−1 reported from the real industrial D. salina production pro-cess in Israel (Sun et al., 2011) which is identical in area size and productivity to the reference case. Furthermore, the overall biomass production cost calculated to be
3 ASSESSMENT OF MICROALGAE BASED PRODUCTION PROCESSES
0 . 0 0 . 1 0 . 2 0 . 3 0 . 4 0 . 5 0 . 6
M a t e r i a l s E n e r g y c o s t T a x
x 1 0 6
Costs (USD a-1 )
M a n p o w e r
U S D k g- 1d w U S D k g- 1ß - c a r o t e n e 1 7 . 1 3 ± 1 . 5 9 3 4 3 . 4 5 ± 3 1 . 9 3
Figure 3.3: Annual operating costs for the production ofβ-carotene with D. salina. Costs are composed of manpower cost, tax, operating costs for energy and raw materials cost.
Error bars account for one standard deviation from the average value based on Monte Carlo simulation to consider the uncertainties of experimental and literature data.
17.13 ± 1.59 USD kg−1dw biomass are almost the same to the value of 17 USD kg−1dw biomass, published for the mentioned production site (Ben-Amotz, 2008). Accordingly, the techno-economic estimates of the process model seem to be realistic, allowing the application of the evaluation concept for the assessment of innovative downstream pro-cesses forβ-carotene production with D. salina in the following chapters of the present thesis.
4
Flocculation as potential
preconcentration step of D. salina
The present chapter is dedicated to the investigation of occulation as possible harvest-ing step within the D. salina downstreamharvest-ing chain. The theoretical feasibility of occu-lation is addressed by the characterization of its physico-chemical properties. Therefore, theoretical and experimental methods established for colloid process engineering are transferred to the algal system. In addition, dierent chemical and physical occulation strategies are experimentally investigated. For the assessment of occulation as innova-tive harvesting approach, dierent assessment criteria are taking into account, e.g. the reusability of separated culture medium or the product yield. Finally, the calculation of operating energy and costs is done to compare all occulation attempts with the conventional harvesting technique centrifugation. This chapter involves methods and results published in Pirwitz et al. (2015a) and Pirwitz et al. (2015b).